cancel
Showing results for 
Search instead for 
Did you mean: 

which process uses these directories y how to control its size?

rbadilla
Level 3
Partner

Hi people:

I'm hunting down the responsible process that are using 52 GB  on /usr/openv/logs/nbrntd  and 220 GB on /usr/openv/var/global/telemetry

:/usr/openv/logs# du -sh *
20K bmrd
1K bmrs
1K bmrsetup
4.8M nbars
4.3M nbatd
404K nbaudit
923M nbdisco
250M nbemm
337K nbevtmgr
5K nbftclnt
346K nbim
246M nbjm
45K nbkms
25M nbpem
43M nbpxyhelper
309M nbrb
22M nbrmms
52G nbrntd
80M nbsl
34M nbstserv
2.8G nbsvcmon
30K nbvault
35K nbwebservice
1K ncfnbbrowse

[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[[

:/usr/openv/var/global/telemetry# du -sh *
1K dataset
1K hostuuid
220G nb_runtime
1.2M scheduling
4.3M systemstats
4.3M tmp
228K upload_cache

Some background data:

Physical host with Solaris 11.3 for X86, 128 GB RAM, one CPU. Currently, running NBU 8.1 in service for 10 years (migrated from Sparc T2 server with older version of NBU)

I used vxlogmgr -d -e "09/25/2016 00:00:00 AM" but didn't find any file to delete... 

So I'm lost finding which process uses those directories and how to control its growing size. ZFS pools are impacted when they are over 90% used. 

Can somebody shed me light over this ?

Thanks a lot in advance.

 

 

 

 

5 REPLIES 5

quebek
Moderator
Moderator
   VIP    Certified

Hey

Check out what are the current settings for nbrtnd:

vxlogcfg -l -p NB -o 387

if diagnostic or debug is set to anything else that 1 - set it to 1 with below

vxlogcfg -a -p NB -o 387 -s DiagnosticLevel=1 -s DebugLevel=1

About telemetry - what is inside this one 220G nb_runtime ?

 

Hi quebek:

About your question on vxlogcfg -l -p NB -o 387, this is the output:

Configuration settings for originator 387, of product 51,216...
LogDirectory = /usr/openv/logs/nbrntd/
DebugLevel = 1
DiagnosticLevel = 6
DynaReloadInSec = 0
LogToStdout = False
LogToStderr = False
LogToOslog = False
RolloverMode = FileSize | LocalTime
LogRecycle = False
MaxLogFileSizeKB = 51200
RolloverPeriodInSeconds = 43200
RolloverAtLocalTime = 0:00
NumberOfLogFiles = 3
OIDNames = nbrntd
AppMsgLogging = ON
L10nLib = /usr/openv/lib/libvxexticu
L10nResource = nbrntd
L10nResourceDir = /usr/openv/resources
SyslogIdent = VRTS-NB
SyslogOpt = 0
SyslogFacility = LOG_LOCAL5
LogFilePermissions = 664


quebek
Moderator
Moderator
   VIP    Certified

Hi

So run this one

vxlogcfg -a -p NB -o 387 -s DiagnosticLevel=1 -s DebugLevel=1

it will decrease the logging level for nbrtnd...

Also you can try to limit logs size by:

"About limiting the size of unified and legacy logs

To limit the size of the NetBackup logs, specify the log size in the Keep logs up to GB option in the NetBackup Administration Console. When the NetBackup log size grows up to this configuration value, the older logs are deleted. To set the log size in GB, select the check box, which lets you select the value in GB from the drop-down list."

This is from https://www.veritas.com/support/en_US/doc/86063237-127664549-0/v106233759-127664549

Nicolai
Moderator
Moderator
Partner    VIP   

vxlogmgr -d  is the preferred option, but if that isn't working go to  /usr/openv/logs/nbrntd and delete the log files, it will have no system impact on Netbackup. If someone zipped a lot of old log files, vxlogmgr will not delete them.

As @quebek has already mentioned, lower the debug/diagnostic level of nbrntd

With regards to /usr/openv/var/global/telemetry, data in this area is information reported back to Veritas, according to link below. I would assume data is safe to delete, try disabling telemetry fist to be on the safe side. Alternative open a support ticket with Veritas.

https://www.veritas.com/support/en_US/doc/133778034-133778053-0/v133387180-133778053

Hi people, I'm apology by having being on silence these days but was waiting to have another session with the final user.

Seem that files over  /usr/openv/var/global/telemetry are somehow part of the catalog backup since; they were running its catalog policy and since takes 4 hours to complete take this screenshot.

2022-06-15 14_50_28-BKP.png

Will check the pathnames on that policy, but if was wizard generated; could be an error from the nbu software (they are using NBU 8.1; since got issues going to 9.0 so they rolled back)?.

Tomorrow with have another session but to upgrade their Solaris 11.3 to 11.4, but I will surely take time to sniff every place to figure out why is so large and the value of those files.